Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proposition A.4

Open
QAlgorithms.Papers.Independent.Support50.P12.proposition_A_4

by Goku · Oct 23, 2026 · Mathlib 0df444a (Lean v4.33.1)

prove2quantum

Let A,BA,BA,B be positive semidefinite operators on a nonzero finite-dimensional complex Hilbert space. Let f:[0,∞)→Rf:[0,\infty)\to\mathbb Rf:[0,∞)→R be continuous on (0,∞)(0,\infty)(0,∞) with signed extended recession slope ω(f)=lim⁡x→∞f(x)/x\omega(f)=\lim_{x\to\infty}f(x)/xω(f)=limx→∞​f(x)/x. Write SfS_fSf​ for the paper's quantum divergence, Qf(A,B)=⟨B1/2,f(LARB−1)B1/2⟩HSQ_f(A,B)=\langle B^{1/2},f(L_AR_{B^{-1}})B^{1/2}\rangle_{\rm HS}Qf​(A,B)=⟨B1/2,f(LA​RB−1​)B1/2⟩HS​, and X0X^0X0 for the support projection. All powers vanish on the kernel, including exponent zero. Traces are representation traces.

Assume Be0B e0Be0 and fff is convex on [0,∞)[0,\infty)[0,∞). Let EB(X)=∑bQbXQbE_B(X)=\sum_bQ_bXQ_bEB​(X)=∑b​Qb​XQb​ be the spectral pinching. Then

Sf(A∥B)≥Sf(EB(A)∥EB(B))=Sf(EB(A)∥B)≥(Tr⁡B)f ⁣(Tr⁡ATr⁡B).S_f(A\Vert B)\ge S_f(E_B(A)\Vert E_B(B))=S_f(E_B(A)\Vert B)\ge(\operatorname{Tr}B)f\!\left(\frac{\operatorname{Tr}A}{\operatorname{Tr}B}\right).Sf​(A∥B)≥Sf​(EB​(A)∥EB​(B))=Sf​(EB​(A)∥B)≥(TrB)f(TrBTrA​).

If A0≤B0A^0\le B^0A0≤B0 and fff is strictly convex, equality in the first inequality is equivalent to AB=BAAB=BAAB=BA; equality in the second is equivalent to EB(A)E_B(A)EB​(A) being a nonnegative multiple of BBB; equality between the extreme terms is equivalent to AAA being a nonnegative multiple of BBB. Both additional hypotheses govern every equality characterization.

Preamble
import Definitions.Def_QAlgorithms_Papers_Independent_Support50_P12
Formal statement
noncomputable section

namespace QAlgorithms.Papers.Independent.Support50.P12

open Filter MeasureTheory Set

open QAlgorithms.Papers.Independent.Information100.P60 (Mat copies)

open QAlgorithms.Papers.Independent.Information100.P30 (MatrixMap)

open QAlgorithms.Papers.Independent.Dependency50.P28 (tr LE logNonnegativeExtended pinch)

open QAlgorithms.Papers.Independent.Dependency50.P21 (channelAdjoint spectralFunction petz relativeEntropy)

open QAlgorithms.Papers.Independent.Frontier18.P17 (hs)

open QAlgorithms.Papers.Independent.Next46.P23 (Supports)

open QAlgorithms.Papers.Independent.Support21.P03 (ExtendedConvex)

open QAlgorithms (spectralProjector)

open QAlgorithms.Learning (IsDistribution)

open QAlgorithms.Papers.Independent.Information100.P04 (absOp)

open scoped BigOperators Matrix ComplexConjugate ComplexOrder ENNReal Topology Kronecker

attribute [local instance] Classical.propDecidable Classical.decEq

open QAlgorithms.Papers.Independent.Information100.P71 (supportPower)

variable {I J K : Type} [Fintype I] [DecidableEq I] [Fintype J] [DecidableEq J]
  [Fintype K] [DecidableEq K]

/-- Proposition A.4, PDF pp.44–45: all inequalities, the pinching identity,
and ALL equality characterizations, each retaining strict convexity and support inclusion. -/
theorem proposition_A_4 [Nonempty I] (X Y : Mat I) (hX : X.PosSemidef) (hY : Y.PosSemidef)
    (hn : Y ≠ 0) (f : ℝ → ℝ) (hf : DivergenceGenerator f) (hc : ConvexOn ℝ (Ici 0) f) :
    div f (pinch Y X) (pinch Y Y) ≤ div f X Y ∧
    div f (pinch Y X) (pinch Y Y) = div f (pinch Y X) Y ∧
    ((tr Y * f (tr X / tr Y) : ℝ) : EReal) ≤ div f (pinch Y X) Y ∧
    (Supports X Y → StrictConvexOn ℝ (Ici 0) f →
      (div f X Y = div f (pinch Y X) Y ↔ Commute X Y) ∧
      (div f (pinch Y X) Y = ((tr Y * f (tr X / tr Y) : ℝ) : EReal) ↔
        ∃ c : ℝ, 0 ≤ c ∧ pinch Y X = (c : ℂ) • Y) ∧
      (div f X Y = ((tr Y * f (tr X / tr Y) : ℝ) : EReal) ↔
        ∃ c : ℝ, 0 ≤ c ∧ X = (c : ℂ) • Y)) := by sorry

end QAlgorithms.Papers.Independent.Support50.P12

end
Source
Quantum f-divergences and error correction, arXiv:1008.2529v6, PDF p. 44, Proposition A.4
Read-back

What the Lean code literally says, in plain math · gpt-6.1-sol

Let III be a finite nonempty index type with decidable equality, let X,Y∈MI(C)X,Y\in M_I(\mathbb C)X,Y∈MI​(C) be positive semidefinite with Y≠0Y\ne0Y=0, and let f:R→Rf:\mathbb R\to\mathbb Rf:R→R be convex on [0,∞)[0,\infty)[0,∞), continuous on (0,∞)(0,\infty)(0,∞), and such that f(x)/xf(x)/xf(x)/x, regarded as extended real, converges at infinity to its totalized limit. Write tr⁡R=Re⁡Tr⁡\operatorname{tr}_{\mathbb R}=\operatorname{Re}\operatorname{Tr}trR​=ReTr, and let

PY(Z)=∑b∈spec⁡YPY,bZPY,b,\mathcal P_Y(Z)=\sum_{b\in\operatorname{spec}Y}P_{Y,b}ZP_{Y,b},PY​(Z)=b∈specY∑​PY,b​ZPY,b​,

where the sum is over distinct eigenvalues, including zero. For positive semidefinite U,VU,VU,V, set

qf(U,V)=∑a∈spec⁡U∑b∈spec⁡Vb f(a/b)Re⁡Tr⁡(PU,aPV,b),q_f(U,V)=\sum_{a\in\operatorname{spec}U} \sum_{b\in\operatorname{spec}V} b\,f(a/b)\operatorname{Re}\operatorname{Tr}(P_{U,a}P_{V,b}),qf​(U,V)=a∈specU∑​b∈specV∑​bf(a/b)ReTr(PU,a​PV,b​),

using a/0=0a/0=0a/0=0, and define Df(U,V)D_f(U,V)Df​(U,V) to be qf(U,V)q_f(U,V)qf​(U,V) when ker⁡V⊆ker⁡U\ker V\subseteq\ker UkerV⊆kerU, and otherwise the totalized extended-real limit of qf(U,V+ε1)q_f(U,V+\varepsilon1)qf​(U,V+ε1) as ε↓0\varepsilon\downarrow0ε↓0. Then

Df(PY(X),PY(Y))≤Df(X,Y),Df(PY(X),PY(Y))=Df(PY(X),Y),tr⁡RY f ⁣(tr⁡RXtr⁡RY)≤Df(PY(X),Y).\begin{aligned} D_f(\mathcal P_Y(X),\mathcal P_Y(Y))&\le D_f(X,Y),\\ D_f(\mathcal P_Y(X),\mathcal P_Y(Y))&=D_f(\mathcal P_Y(X),Y),\\ \operatorname{tr}_{\mathbb R}Y\, f\!\left(\frac{\operatorname{tr}_{\mathbb R}X}{\operatorname{tr}_{\mathbb R}Y}\right) &\le D_f(\mathcal P_Y(X),Y). \end{aligned}Df​(PY​(X),PY​(Y))Df​(PY​(X),PY​(Y))trR​Yf(trR​YtrR​X​)​≤Df​(X,Y),=Df​(PY​(X),Y),≤Df​(PY​(X),Y).​

Moreover, if ker⁡Y⊆ker⁡X\ker Y\subseteq\ker XkerY⊆kerX and fff is strictly convex on [0,∞)[0,\infty)[0,∞), then

Df(X,Y)=Df(PY(X),Y)  ⟺  XY=YX,Df(PY(X),Y)=tr⁡RY f ⁣(tr⁡RXtr⁡RY)  ⟺  ∃c≥0, PY(X)=cY,Df(X,Y)=tr⁡RY f ⁣(tr⁡RXtr⁡RY)  ⟺  ∃c≥0, X=cY.\begin{aligned} D_f(X,Y)=D_f(\mathcal P_Y(X),Y) &\iff XY=YX,\\ D_f(\mathcal P_Y(X),Y)= \operatorname{tr}_{\mathbb R}Y\,f\!\left(\frac{\operatorname{tr}_{\mathbb R}X}{\operatorname{tr}_{\mathbb R}Y}\right) &\iff \exists c\ge0,\ \mathcal P_Y(X)=cY,\\ D_f(X,Y)= \operatorname{tr}_{\mathbb R}Y\,f\!\left(\frac{\operatorname{tr}_{\mathbb R}X}{\operatorname{tr}_{\mathbb R}Y}\right) &\iff \exists c\ge0,\ X=cY. \end{aligned}Df​(X,Y)=Df​(PY​(X),Y)Df​(PY​(X),Y)=trR​Yf(trR​YtrR​X​)Df​(X,Y)=trR​Yf(trR​YtrR​X​)​⟺XY=YX,⟺∃c≥0, PY​(X)=cY,⟺∃c≥0, X=cY.​

Neither matrix is required to have trace 111, and X=0X=0X=0 is allowed.

Human review
  • Endorsed by Shuze Chen · Oct 24, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Goku · Oct 24, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me